DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Status of the Claims
Claims 1-4, and 6-20 are currently pending in the present application, with claims 1, 13, and 17 being independent.
Response to Amendments / Arguments
Applicant’s arguments, see Pg. 16, filed 12/08/2025, with respect to claim 11 have been fully considered and are persuasive. The objection of claim 11 has been withdrawn.
Applicant’s arguments, see Pg. 16-23, filed 12/08/2025, with respect to the rejection(s) of claim(s) 1-3, 6-8, 11, and 17 under 35 U.S.C. § 102 have been fully considered and are persuasive. Therefore, the rejection has been withdrawn. However, upon further consideration, a new ground(s) of rejection is made in view of newly found prior art.
Applicant’s arguments with respect to claim(s) 4 and 6 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Regarding the remaining arguments: Applicant argues with respect to the amended claim language, which is fully addressed in the prior art rejections set forth below.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1-4, 8, 11, and 17-18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Chang et al. "New adaptive color quantization method based on self-organizing maps." IEEE transactions on neural networks 16, no. 1 (2005): 237-249, hereinafter referred to as “Chang”, in view of Giraud et al. "Superpixel-based color transfer." In 2017 IEEE International Conference on Image Processing (ICIP), pp. 700-704. IEEE, 2017, hereinafter referred to as “Giraud”.
Regarding claim 1, Chang discloses a computer-implemented method comprising:
accessing a color palette comprising colors (Pg. 239; new image…original color image);
mapping colors to a three-dimensional (3-D) color space by outputting the colors at locations on the 3-D color space (Pg. 239, Left Column, Step 2); A new image pixel value x(t) = [r(t), g(t), b(t)] ^T);
generating a distribution of points on the 3-D color space by outputting the distribution of points at corresponding locations on the 3-D color space (Embodiment 1 (SOM): Fig. 2 and Pg. 239, Left Column, Step 3); The input x(t) is compared to the weight wi(t) of each neuron i simultaneously, by the square of Euclidean distance in red, green, and blue (RGB) space defined as follows: (1). Embodiment 2 (FS-SOM): Pg. 240, Left Column, 1); initialize the neurons of FS-SOM with uniformly distributed weight vectors on the major diagonal of the input space, i.e. (5) where N is the number of neurons);
organizing via a self-organizing map algorithm, the distribution of points (Embodiment 1 (SOM): Pg. 239, Left Column, Par. 1; Let N be the total number of neurons and wi(t) ∈ R^3 be the weight vector of the ith neuron at epoch time t. Embodiment 2 (FS-SOM): Pg. 240, Left Column, 1); initialize the neurons of FS-SOM with uniformly distributed weight vectors on the major diagonal of the input space, i.e. (5) where N is the number of neurons) to discrete locations on the 3-D color space to output an organized distribution of points on the 3-D color space (Embodiment 1 (SOM): Pg. 239, Left Column, Step 3); the weight wi(t) of each neuron i…in red, green, and blue (RGB) space…Embodiment 2 (FS-SOM): Pg. 240, Left Column, 2); a global butterfly permutation sequence is proposed to present the spatially correlated input data from a multidimensional coordinate system), wherein organizing the distribution of points is based on a relationship between the corresponding locations of the organized distribution of points on the 3-D color space and the locations of the colors on the 3-D color space (Embodiment 1 (SOM): Pg. 239, Left-Right Column; the neurons training to obtain the color palette (Steps 2-5)…Step 3) …The input x(t) is compared to the weight wi(t) of each neuron I simultaneously, by the square of Euclidean distance in red, green, and blue space defined as follows: (1)-(4). Embodiment 2 (FS-SOM): Fig. 7-8 and Pg. 243, Section IV, Right Column; Neighboring neurons linked by solid lines represent adjacent entries in the color palette. The absence of long links indicates a good color-ordering palette…By superimpose the topological maps of Fig. 8 on the color distribution of the original images in Fig. 7);
computing a first set of Euclidian distances between a first location of the locations corresponding to a first color of the colors to the discrete locations of the organized distribution of points (Embodiment 1 (SOM): Pg. 239, Left Column, Step 3); The input tristimulus x(t) is compared to the weight wi(t) of each neuron I simultaneously, by the square of Euclidian distance in red, green, and blue (RGB) space…||x(t) - wi(t)||^2…Embodiment 2 (FS-SOM): Pg. 240, Right Column, 3)-4); adopting the standard SOM best marching unit search with F(ui) = 1…H(i, c, m) is a tapering neighborhood function which is a hallmark of the SOM algorithms. This neighborhood taper alters the update of each neuron, I within the neighborhood of winning neuron, c by a different amount at discrete time interval, m depending on the distance between weight vectors wi and wc…In (7) and (8) the use of a frequency sensitive learning rate function implies that the training time for each individual neuron, rather than the training time of the network, is used to modulate the distance metric for updating);
associating the first color with a first point of the organized distribution points based on a shortest Euclidian distance of the first set of Euclidian distances (Embodiment 1 (SOM): Pg. 239, Left Column, Step 3); The winning neuron c for this input is the one with the smallest distance, i.e., (2)…Right Column, Step 6); For every pixel value in the original color image, the best representative color is ground from the color map based on the same Euclidean distance metric of (1). Embodiment 2 (FS-SOM): Pg. 240, Right Column, 3)-4); adopting the standard SOM best marching unit search with F(ui) = 1…);
computing a second set of Euclidian distances between a second location of the locations corresponding to a second color of the colors to the discrete locations of the organized distribution of points (Embodiment 1 (SOM): Pg. 239, Left Column, Step 3); The input tristimulus x(t) is compared to the weight wi(t) of each neuron I simultaneously, by the square of Euclidian distance in red, green, and blue (RGB) space…||x(t) - wi(t)||^2…Pg. 239, Right Column, Step 4); The epoch time t is increased by one…The process is repeated from Step 2… Embodiment 2 (FS-SOM): Pg. 240, Right Column, 3)-4); adopting the standard SOM best marching unit search with F(ui) = 1…); and
associating the second color with a second point of the organized distribution points (Embodiment 1 (SOM): Pg. 239, Left Column, Step 3); The winning neuron c for this input is the one with the smallest distance, i.e., (2)…Right Column, Step 6); For every pixel value in the original color image, the best representative color is ground from the color map based on the same Euclidean distance metric of (1)... Pg. 239, Right Column, Step 4); The epoch time t is increased by one…The process is repeated from Step 2… Embodiment 2 (FS-SOM): Pg. 240, Right Column, 3)-4); adopting the standard SOM best marching unit search with F(ui) = 1…),
sorting the colors based on the discrete locations, on the 3-D color space, of the organized distribution of points associated with the colors to output the color palette comprising the colors as sorted (Fig. 7-8 and Pg. 243, Section IV, Right Column; Neighboring neurons linked by solid lines represent adjacent entries in the color palette. The absence of long links indicates a good color-ordering palette
and causing display of the color palette comprising the colors as sorted (Fig. 13-14).
Chang does not disclose associating the colors with different points of the organized distribution of points, such that no two colors of the colors are associated with a same point of the organized distribution of points, and associating the second color with a second point of the organized distribution points that is not already associated with the first color.
In the same art of associating/matching color-representative elements using a nearest-neighbor distance criterion, Giraud discloses associating the colors with different points of the organized distribution of points, such that no two colors of the colors are associated with a same point of the organized distribution of points (Pg. 701, Right Column, Section 2.2, Par. 2; propose to constrain the ANN search and to restrict the number of associations to the same element. To do so, we set a parameter E that defines the maximum number of selection of the same superpixel…E = 1…during the iterative process, a superpixel Ai can be assigned to a superpixel Bk, only if less than E superpixels in A are already assigned to Bk. If Bk is already matched by E elements in A, one superpixel Aj assigned to Bk must be sent to another superpixel in B to allow Ai to match Bk…), and associating the second color with a second point of the organized distribution points that is not already associated with the first color (Pg. 701, Right Column, Section 2.2, Par. 2; propose to constrain the ANN search and to restrict the number of associations to the same element. To do so, we set a parameter E that defines the maximum number of selection of the same superpixel…E = 1…during the iterative process, a superpixel Ai can be assigned to a superpixel Bk, only if less than E superpixels in A are already assigned to Bk. If Bk is already matched by E elements in A, one superpixel Aj assigned to Bk must be sent to another superpixel in B to allow Ai to match Bk…).
It would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to incorporate Giraud’s constraint on match diversity into Chang’s nearest-distance (best-matching) color to node association when forming a palette. Chang’s baseline nearest-neighbor assignment can predictably produce multiple colors mapping to the same representative point that could reduce palette diversity and stable ordering, therefore, applying Giraud’s one-to-one assignment constraint would be a routine optimization in the art of color-space mapping to ensure unique nearest-neighbor assignment, yielding predictable results of better distributed set of representative points for subsequent palette sorting and display.
Regarding claim 2, Chang in view of Giraud discloses the computer-implemented method of claim 1, and further discloses wherein the distribution of points is randomly generated before organization by the self-organizing map algorithm (Chang Embodiment 1: Pg. 239, Left Column, Step 1); the weights of the neurons, wi(0)…are initialized according to a selected initialization criterion. Embodiment 2: Pg. 240, Left Column, 2); Butterfly Permutation for Input Randomization).
Chang and Giraud are combined for the reasons set forth above with respect to claim 1.
Regarding claim 3, Chang in view of Giraud discloses the computer-implemented method of claim 1, and further discloses wherein the distribution of points on the 3-D color space is organized by the self-organizing map algorithm by iteratively optimizing the corresponding locations of the distribution of points based on the relationship between the corresponding locations on the 3-D color space and the locations of the colors on the 3-D color space (Chang Embodiment 1 (SOM): Pg. 239, Left Column, Step 3); The input tristimulus x(t) is compared to the weight wi(t) of each neuron I simultaneously, by the square of Euclidian distance in red, green, and blue (RGB) space…||x(t) - wi(t)||^2…Pg. 239, Right Column, Step 4); The epoch time t is increased by one…The process is repeated from Step 2 until the average difference in neuron weights between two successive iterations converges to a small value…Embodiment 2 (FS-SOM): Pg. 240, Right Column, 3)-4); adopting the standard SOM best marching unit search with F(ui) = 1…).
Chang and Giraud are combined for the reasons set forth above with respect to claim 1.
Regarding claim 4, Chang in view of Giraud discloses the computer-implemented method of claim 1, and further discloses wherein the distribution of points comprises an equal number of points to the colors (Chang Pg. 242, Section III.B, Par. 1; When SOM is used for CQ, the color palette is formed from the final weights wi of the neurons. Hence, the maximum number of colors that can appear in the reconstructed image is equal to the number of neurons N), and wherein associating the colors with different points of the organized distribution of points occurs after organizing the distribution of points to output the organized distribution of points (Chang Pg. 239, Section II; three main phase…the initial network setup (Step 1_, the neurons training to obtain the color palette (Steps 2-5) and the quantizer mapping to generate the pixel indices (Step 6)…Step 5) colormap design has been completed…Step 6) the best representative color is found from the color map based on the same Euclidian distance metric…)
Chang does not disclose such that all points of the organized distribution of points are associated with a different color of the colors after organization via the self-organizing map algorithm.
In the same art of associating/matching color-representative elements using a nearest-neighbor distance criterion, Giraud discloses such that all points of the organized distribution of points are associated with a different color of the colors after organization via the self-organizing map algorithm (Pg. 701, Right Column, Section 2.2, Par. 2; propose to constrain the ANN search and to restrict the number of associations to the same element. To do so, we set a parameter E that defines the maximum number of selection of the same superpixel…E = 1…during the iterative process, a superpixel Ai can be assigned to a superpixel Bk, only if less than E superpixels in A are already assigned to Bk. If Bk is already matched by E elements in A, one superpixel Aj assigned to Bk must be sent to another superpixel in B to allow Ai to match Bk…).
Chang and Giraud are combined for the reasons set forth above with respect to claim 1.
Regarding claim 8, Chang in view of Giraud discloses the computer-implemented method of claim 1, and further discloses further comprising: adding one or more corresponding colors represented by the discrete locations of the organized distribution of points to the color palette (Chang Pg. 239, Left Column, Par. 1; the neurons training to obtain the color palette (Steps 2-5)…Pg. 242, Left Column, Section III.B; When SOM is used for CQ, the color palette is formed from the final weights wi of the neurons. Pg. 243, Section IV, Right Column; Neighboring neurons linked by solid lines represent adjacent entries in the color palette. The absence of long links indicates a good color-ordering palette…By superimpose the topological maps of Fig. 8 on the color distribution of the original images in Fig. 7);
and display of the color palette with the one or more corresponding colors (Fig. 13-14).
Chang and Giraud are combined for the reasons set forth above with respect to claim 1.
Regarding claim 11, Chang in view of Giraud discloses the computer-implemented method of claim 1, and further discloses sorting the colors to output the color palette as a two-dimensional (2-D) color palette (Chang Fig. 7-8 and Pg. 243, Section IV, Right Column; Neighboring neurons linked by solid lines represent adjacent entries in the color palette. The absence of long links indicates a good color-ordering palette);
and causing display of the color palette as the 2-D color palette (Fig. 13-14).
Chang and Giraud are combined for the reasons set forth above with respect to claim 1.
Regarding claim 17, Chang discloses a computing system comprising:
a processor; and a non-transitory computer-readable medium having stored thereon instructions that when executed by the processor, cause the processor to perform operations (Section IV, Pg. 245, Right Column, Par. 2; All the simulations were run on a Personal Computer equipped with a Pentium Pro. 2.1 GHz Processor and 512 Mbytes of system memory) including:
accessing a two-dimensional (2-D) color palette comprising an array of colors (Pg. 239; new image…original color image),
mapping the array of colors to a three-dimensional (3-D) color space by outputting the array of colors at locations on the 3-D color space (Pg. 239, Left Column, Step 2); A new image pixel value x(t) = [r(t), g(t), b(t)] ^T),
generating a distribution of points on the 3-D color space by outputting the distribution of points at corresponding locations on the 3-D color space (Embodiment 1 (SOM): Fig. 2 and Pg. 239, Left Column, Step 3); The input x(t) is compared to the weight wi(t) of each neuron i simultaneously, by the square of Euclidean distance in red, green, and blue (RGB) space defined as follows: (1). Embodiment 2 (FS-SOM): Pg. 240, Left Column, 1); initialize the neurons of FS-SOM with uniformly distributed weight vectors on the major diagonal of the input space, i.e. (5) where N is the number of neurons);
organizing via a self-organizing map algorithm, the distribution of points (Embodiment 1 (SOM): Pg. 239, Left Column, Par. 1; Let N be the total number of neurons and wi(t) ∈ R^3 be the weight vector of the ith neuron at epoch time t. Embodiment 2 (FS-SOM): Pg. 240, Left Column, 1); initialize the neurons of FS-SOM with uniformly distributed weight vectors on the major diagonal of the input space, i.e. (5) where N is the number of neurons) to discrete locations on the 3-D color space to output an organized distribution of points on the 3-D color space (Embodiment 1 (SOM): Pg. 239, Left Column, Step 3); the weight wi(t) of each neuron i…in red, green, and blue (RGB) space…Embodiment 2 (FS-SOM): Pg. 240, Left Column, 2); a global butterfly permutation sequence is proposed to present the spatially correlated input data from a multidimensional coordinate system), wherein organizing the distribution of points is based on a relationship between the corresponding locations of the organized distribution of points on the 3-D color space and the locations of the colors on the 3-D color space (Embodiment 1 (SOM): Pg. 239, Left-Right Column; the neurons training to obtain the color palette (Steps 2-5)…Step 3) …The input x(t) is compared to the weight wi(t) of each neuron I simultaneously, by the square of Euclidean distance in red, green, and blue space defined as follows: (1)-(4). Embodiment 2 (FS-SOM): Fig. 7-8 and Pg. 243, Section IV, Right Column; Neighboring neurons linked by solid lines represent adjacent entries in the color palette. The absence of long links indicates a good color-ordering palette…By superimpose the topological maps of Fig. 8 on the color distribution of the original images in Fig. 7);
computing a first set of Euclidian distances between a first location of the locations corresponding to a first color of the colors to the discrete locations of the organized distribution of points (Embodiment 1 (SOM): Pg. 239, Left Column, Step 3); The input tristimulus x(t) is compared to the weight wi(t) of each neuron I simultaneously, by the square of Euclidian distance in red, green, and blue (RGB) space…||x(t) - wi(t)||^2…Embodiment 2 (FS-SOM): Pg. 240, Right Column, 3)-4); adopting the standard SOM best marching unit search with F(ui) = 1…H(i, c, m) is a tapering neighborhood function which is a hallmark of the SOM algorithms. This neighborhood taper alters the update of each neuron, I within the neighborhood of winning neuron, c by a different amount at discrete time interval, m depending on the distance between weight vectors wi and wc…In (7) and (8) the use of a frequency sensitive learning rate function implies that the training time for each individual neuron, rather than the training time of the network, is used to modulate the distance metric for updating);
associating the first color with a first point of the organized distribution points based on a shortest Euclidian distance of the first set of Euclidian distances (Embodiment 1 (SOM): Pg. 239, Left Column, Step 3); The winning neuron c for this input is the one with the smallest distance, i.e., (2)…Right Column, Step 6); For every pixel value in the original color image, the best representative color is ground from the color map based on the same Euclidean distance metric of (1). Embodiment 2 (FS-SOM): Pg. 240, Right Column, 3)-4); adopting the standard SOM best marching unit search with F(ui) = 1…);
computing a second set of Euclidian distances between a second location of the locations corresponding to a second color of the array of colors to the discrete locations of the organized distribution of points (Embodiment 1 (SOM): Pg. 239, Left Column, Step 3); The input tristimulus x(t) is compared to the weight wi(t) of each neuron I simultaneously, by the square of Euclidian distance in red, green, and blue (RGB) space…||x(t) - wi(t)||^2…Pg. 239, Right Column, Step 4); The epoch time t is increased by one…The process is repeated from Step 2… Embodiment 2 (FS-SOM): Pg. 240, Right Column, 3)-4); adopting the standard SOM best marching unit search with F(ui) = 1…); and
associating the second color with a second point of the organized distribution points (Embodiment 1 (SOM): Pg. 239, Left Column, Step 3); The winning neuron c for this input is the one with the smallest distance, i.e., (2)…Right Column, Step 6); For every pixel value in the original color image, the best representative color is ground from the color map based on the same Euclidean distance metric of (1)... Pg. 239, Right Column, Step 4); The epoch time t is increased by one…The process is repeated from Step 2… Embodiment 2 (FS-SOM): Pg. 240, Right Column, 3)-4); adopting the standard SOM best marching unit search with F(ui) = 1…),
sorting the colors based on the discrete locations, on the 3-D color space, of the organized distribution of points associated with the colors to output the color palette comprising the colors as sorted (Fig. 7-8 and Pg. 243, Section IV, Right Column; Neighboring neurons linked by solid lines represent adjacent entries in the color palette. The absence of long links indicates a good color-ordering palette
and causing display of the color palette comprising the colors as sorted (Fig. 13-14).
Chang does not disclose associating the array of colors with different points of the organized distribution of points, such that no two colors of the array of colors are associated with a same point of the organized distribution of points, and associating the second color with a second point of the organized distribution points that is not already associated with the first color.
In the same art of associating/matching color-representative elements using a nearest-neighbor distance criterion, Giraud discloses associating the colors with different points of the organized distribution of points, such that no two colors of the colors are associated with a same point of the organized distribution of points (Pg. 701, Right Column, Section 2.2, Par. 2; propose to constrain the ANN search and to restrict the number of associations to the same element. To do so, we set a parameter E that defines the maximum number of selection of the same superpixel…E = 1…during the iterative process, a superpixel Ai can be assigned to a superpixel Bk, only if less than E superpixels in A are already assigned to Bk. If Bk is already matched by E elements in A, one superpixel Aj assigned to Bk must be sent to another superpixel in B to allow Ai to match Bk…), and associating the second color with a second point of the organized distribution points that is not already associated with the first color (Pg. 701, Right Column, Section 2.2, Par. 2; propose to constrain the ANN search and to restrict the number of associations to the same element. To do so, we set a parameter E that defines the maximum number of selection of the same superpixel…E = 1…during the iterative process, a superpixel Ai can be assigned to a superpixel Bk, only if less than E superpixels in A are already assigned to Bk. If Bk is already matched by E elements in A, one superpixel Aj assigned to Bk must be sent to another superpixel in B to allow Ai to match Bk…).
It would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to incorporate Giraud’s constraint on match diversity into Chang’s nearest-distance (best-matching) color to node association when forming a palette. Chang’s baseline nearest-neighbor assignment can predictably produce multiple colors mapping to the same representative point that could reduce palette diversity and stable ordering, therefore, applying Giraud’s one-to-one assignment constraint would be a routine optimization in the art of color-space mapping to ensure unique nearest-neighbor assignment, yielding predictable results of better distributed set of representative points for subsequent palette sorting and display.
Regarding claim 18, Chang in view of Giraud discloses the computing system of claim 17, and further discloses wherein the distribution of points comprises an equal number of points to the colors (Chang Pg. 242, Section III.B, Par. 1; When SOM is used for CQ, the color palette is formed from the final weights wi of the neurons. Hence, the maximum number of colors that can appear in the reconstructed image is equal to the number of neurons N), and wherein associating the colors with different points of the organized distribution of points occurs after organizing the distribution of points to output the organized distribution of points (Chang Pg. 239, Section II; three main phase…the initial network setup (Step 1_, the neurons training to obtain the color palette (Steps 2-5) and the quantizer mapping to generate the pixel indices (Step 6)…Step 5) colormap design has been completed…Step 6) the best representative color is found from the color map based on the same Euclidian distance metric…).
Chang does not disclose such that all points of the organized distribution of points are associated with a different color of the colors after organization via the self-organizing map algorithm.
In the same art of associating/matching color-representative elements using a nearest-neighbor distance criterion, Giraud discloses such that all points of the organized distribution of points are associated with a different color of the colors after organization via the self-organizing map algorithm (Pg. 701, Right Column, Section 2.2, Par. 2; propose to constrain the ANN search and to restrict the number of associations to the same element. To do so, we set a parameter E that defines the maximum number of selection of the same superpixel…E = 1…during the iterative process, a superpixel Ai can be assigned to a superpixel Bk, only if less than E superpixels in A are already assigned to Bk. If Bk is already matched by E elements in A, one superpixel Aj assigned to Bk must be sent to another superpixel in B to allow Ai to match Bk…).
Chang and Giraud are combined for the reasons set forth above with respect to claim 17.
Claim(s) 6-7, and 19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Chang et al. "New adaptive color quantization method based on self-organizing maps." IEEE transactions on neural networks 16, no. 1 (2005): 237-249, hereinafter referred to as “Chang”, in view of Giraud et al. "Superpixel-based color transfer." In 2017 IEEE International Conference on Image Processing (ICIP), pp. 700-704. IEEE, 2017, hereinafter referred to as “Giraud”, and in further view of Zhang et al. "Color clustering using self-organizing maps." In 2007 International Conference on Wavelet Analysis and Pattern Recognition, vol. 3, pp. 986-989. IEEE, 2007, hereinafter referred to as “Zhang”.
Regarding claim 6, Chang in view of Giraud discloses the computer-implemented method of claim 1, and further discloses (Chang Pg. 239, Section II; three main phase…the initial network setup (Step 1_, the neurons training to obtain the color palette (Steps 2-5) and the quantizer mapping to generate the pixel indices (Step 6)…Step 5) colormap design has been completed…Step 6)the best representative color is found from the color map based on the same Euclidian distance metric…),
Chang does not disclose wherein the distribution of points comprises a number of points that is greater than the colors, associating the colors with different points and such that all of the colors are associated with a different point of a subset of the organized distribution of points after organization via the self-organizing map algorithm.
In the same art of associating/matching color-representative elements using a nearest-neighbor distance criterion, Giraud discloses such that all of the colors are associated with a different point of a subset of the organized distribution of points after organization via the self-organizing map algorithm (Pg. 701, Right Column, Section 2.2, Par. 2; propose to constrain the ANN search and to restrict the number of associations to the same element. To do so, we set a parameter E that defines the maximum number of selection of the same superpixel…E = 1…during the iterative process, a superpixel Ai can be assigned to a superpixel Bk, only if less than E superpixels in A are already assigned to Bk. If Bk is already matched by E elements in A, one superpixel Aj assigned to Bk must be sent to another superpixel in B to allow Ai to match Bk…).
Chang and Giraud are combined for the reasons set forth above with respect to claim 1.
Chang in view of Giraud does not appear to explicitly disclose wherein the distribution of points comprises a number of points that is greater than the colors.
In the same art of color-palette generation/ordering via color space clustering and nearest neighbor assignment, Zhang discloses wherein the distribution of points comprises a number of points that is greater than the colors (Pg. 988, Section 3, Right column; set the size of SOM to 16*16…The image colors are reiteratively used to train the network…each color point is cyclically chosen from the data set, and presented to all neurons on the map simultaneously. Examiner's note: Because each color point is presented to all neurons simultaneously, for any given color point, there exists a plurality of neurons available to be matched, therefore there are more neurons (points) than colors (input color point)).
It would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to incorporate Zhang’s teaching of configuring the SOM with more neurons (points) than input colors into Chang and Giraud’s SOM-based palette construction system. A larger codebook is a routine way to increase representative resolution and reduce quantization error. When Chang and Giraud’s combined system has match-diversity constraint that prevents reuse of the same representative, a larger codebook of candidate neurons is a routine way to increase representative resolution, allowing the system to select a near-optimal alternative rather than forcing multiple inputs into the same representative or accepting a poor match, providing predictable results in improved reconstruction/palette fidelity and output consistency across varying palettes/distributions.
Regarding claim 7, Chang in view of Giraud and in further view of Zhang discloses the computer-implemented method of claim 6, but Chang in view of Giraud does not disclose wherein each point in the organized distribution of points that is not associated with any color of the colors is not used for displaying of the colors as sorted.
In the same art of color-palette generation/ordering via color space clustering and nearest neighbor assignment, Zhang discloses wherein each point in the organized distribution of points that is not associated with any color of the colors is not used for displaying of the colors as sorted (Pg. 988, Section 3, Right column and Fig. 3; The next color point is presented to the network at time t+1. The new winning neuron is chosen by repeating the procedure from step 2 until all iterations have been made. Examiner's note: only the winning neurons are associated with a color in the set of colors, and used for displaying in Fig. 3).
It would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to incorporate Zhang’s display of only selected palette entries into Chang and Giraud’s SOM-based palette construction system. Displaying unassigned points provides no color information and would unnecessarily clutter the palette output/display, therefore the combination is a predictable UI optimization to omit any points to reduce visual clutter and improve user experience.
Regarding claim 19, Chang in view of Giraud discloses the computing system of claim 17, and further discloses wherein associating the array of colors with the different points of the organized distribution of points occurs after organizing the distribution of points to output the organized distribution of points (Chang Pg. 239, Section II; three main phase…the initial network setup (Step 1_, the neurons training to obtain the color palette (Steps 2-5) and the quantizer mapping to generate the pixel indices (Step 6)…Step 5) colormap design has been completed…Step 6)the best representative color is found from the color map based on the same Euclidian distance metric…),
adding one or more corresponding colors represented by the discrete locations of the organized distribution of points (Chang Pg. 239, Left Column, Par. 1; the neurons training to obtain the color palette (Steps 2-5) …Pg. 242, Left Column, Section III.B; When SOM is used for CQ, the color palette is formed from the final weights wi of the neurons. Pg. 243, Section IV, Right Column; Neighboring neurons linked by solid lines represent adjacent entries in the color palette. The absence of long links indicates a good color-ordering palette…By superimpose the topological maps of Fig. 8 on the color distribution of the original images in Fig. 7),
causing display of the 2-D color palette with each corresponding color of each point of the organized distribution of points (Fig. 13-14)
Chang does not disclose wherein the distribution of points comprises a number of points that is greater than colors in the array of colors, such that all of the colors of the array of colors are associated with a different point of a subset of the organized distribution of points after organization via the self-organizing map algorithm, organized distribution of points that are not associated with any color of the array of colors, and organized distribution of points that is not associated with any color in the array of colors.
In the same art of associating/matching color-representative elements using a nearest-neighbor distance criterion, Giraud discloses such that all of the colors of the array of colors are associated with a different point of a subset of the organized distribution of points after organization via the self-organizing map algorithm, organized distribution of points that are not associated with any color of the array of colors, and organized distribution of points that is not associated with any color in the array of colors (Pg. 701, Right Column, Section 2.2, Par. 2; propose to constrain the ANN search and to restrict the number of associations to the same element. To do so, we set a parameter E that defines the maximum number of selection of the same superpixel…E = 1…during the iterative process, a superpixel Ai can be assigned to a superpixel Bk, only if less than E superpixels in A are already assigned to Bk. If Bk is already matched by E elements in A, one superpixel Aj assigned to Bk must be sent to another superpixel in B to allow Ai to match Bk…).
Chang and Giraud are combined for the reasons set forth above with respect to claim 1.
Chang in view of Giraud does not appear to explicitly disclose wherein the distribution of points comprises a number of points that is greater than colors in the array of colors
In the same art of color-palette generation/ordering via color space clustering and nearest neighbor assignment, Zhang discloses wherein the distribution of points comprises a number of points that is greater than colors in the array of colors (Pg. 988, Section 3, Right column; set the size of SOM to 16*16…The image colors are reiteratively used to train the network…each color point is cyclically chosen from the data set, and presented to all neurons on the map simultaneously. Examiner's note: Because each color point is presented to all neurons simultaneously, for any given color point, there exists a plurality of neurons available to be matched, therefore there are more neurons (points) than colors (input color point)).
Chang, Giraud, and Zhang are combined for the reasons set forth above with respect to claim 6.
Claim(s) 9 and 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Chang et al. "New adaptive color quantization method based on self-organizing maps." IEEE transactions on neural networks 16, no. 1 (2005): 237-249, hereinafter referred to as “Chang”, in view of Giraud et al. "Superpixel-based color transfer." In 2017 IEEE International Conference on Image Processing (ICIP), pp. 700-704. IEEE, 2017, hereinafter referred to as “Giraud”, and in further view of Xiao et al. “Self-organizing map-based color palette for high-dynamic range texture compression”. Neural Comput & Applic 21, 639-647 (2012), pages 640-647, hereinafter referred to as “Xiao”.
Regarding claim 9, Chang in view of Giraud discloses the computer-implemented method of claim 1, but does not disclose wherein causing display of the color palette further comprises:
causing display of an adjustable parameter, wherein the adjustable parameter generates linear-interpolated colors by linear-interpolating between the colors as sorted and corresponding colors represented by the discrete locations of the organized distribution of points associated with the colors; receiving a change to the adjustable parameter; and causing display of the linear-interpolated colors based on the change to the adjustable parameter.
In the same art of SOM-based training for color organization, Xiao discloses wherein causing display of the color palette further comprises:
causing display of an adjustable parameter (Pg. 642, Section 3, Formulas (6) and (7) definition of p and q as normalized coordinates. Examiner's note: p and q are adjustable parameters because they determine interpolation weights, adjusting p and q changes the displayed colors as seen in Figure 4), wherein the adjustable parameter generates linear-interpolated colors by linear-interpolating between the colors as sorted and corresponding colors represented by the discrete locations of the organized distribution of points associated with the colors (Pg. 642, Section 3; perform linear interpolation along the horizontal direction to get two intermediate color values, given by Formula (8) and (9)…linear interpolation along the vertical direction, given by Formula (10));
receiving a change to the adjustable parameter (Formulas (6) and (7) p and q. Examiner's note: changing p and q changes which interpolated value is computed, Fig. 3b and 4 showing how changing coordinates values (and thus p and q) leads to new interpolated palette colors);
and causing display of the linear-interpolated colors based on the change to the adjustable parameter (Fig. 4 shows a 16x16 palette expanded to a 256x256 palette by interpolation)
It would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to combine Xiao’s linear interpolation method for virtual palettes with Chang and Giraud’s SOM-based color sorting technique. Doing so provides the predictable benefit of smoother transitions and high-resolution palettes, with interpolation being a well-known and common technique in image processing.
Regarding claim 20, Chang in view of Giraud discloses the computing system of claim 17, but does not disclose wherein causing display of the 2-D color palette further comprises:
causing display of an adjustable parameter, wherein the adjustable parameter generates linear-interpolated colors by linear-interpolating between the array of colors as sorted and corresponding colors represented by the discrete locations of the organized distribution of points associated with the array of colors; receiving a change to the adjustable parameter; and causing display of the linear-interpolated colors based on the change to the adjustable parameter
In the same art of SOM-based training for color organization, Xiao discloses wherein causing display of the 2-D color palette further comprises:
causing display of an adjustable parameter (Pg. 642, Section 3, Formulas (6) and (7) definition of p and q as normalized coordinates. Examiner's note: p and q are adjustable parameters because they determine interpolation weights, adjusting p and q changes the displayed colors as seen in Figure 4), wherein the adjustable parameter generates linear-interpolated colors by linear-interpolating between the array of colors as sorted and corresponding colors represented by the discrete locations of the organized distribution of points associated with the array of colors (Pg. 642, Section 3; perform linear interpolation along the horizontal direction to get two intermediate color values, given by Formula (8) and (9)…linear interpolation along the vertical direction, given by Formula (10));
receiving a change to the adjustable parameter (Formulas (6) and (7) p and q. Examiner's note: changing p and q changes which interpolated value is computed, Fig. 3b and 4 showing how changing coordinates values (and thus p and q) leads to new interpolated palette colors);
and causing display of the linear-interpolated colors based on the change to the adjustable parameter (Fig. 4 shows a 16x16 palette expanded to a 256x256 palette by interpolation)
It would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to combine Xiao’s linear interpolation method for virtual palettes with Chang and Giraud’s SOM-based color sorting technique. Doing so provides the predictable benefit of smoother transitions and high-resolution palettes, with interpolation being a well-known and common technique in image processing.
Claim(s) 10, and 12-14 is/are rejected under 35 U.S.C. 103 as being unpatentable over Chang et al. "New adaptive color quantization method based on self-organizing maps." IEEE transactions on neural networks 16, no. 1 (2005): 237-249, hereinafter referred to as “Chang”, in view of Giraud et al. "Superpixel-based color transfer." In 2017 IEEE International Conference on Image Processing (ICIP), pp. 700-704. IEEE, 2017, hereinafter referred to as “Giraud”, and in further view of obvious design choice (MPEP § 2144.04).
Regarding claim 10, Chang in view of Giraud discloses the computer implemented method of claim 1, and further discloses sorting the colors to output the color palette (Chang Fig. 7-8 and Pg. 243, Section IV, Right Column; Neighboring neurons linked by solid lines represent adjacent entries in the color palette. The absence of long links indicates a good color-ordering palette),
and causing display of the color palette (Chang Fig. 13-14).
Chang in view of Giraud does not appear to explicitly disclose sorting the colors to output the color palette as a one-dimensional (1-D) color palette, and causing display of the color palette as the 1-D color palette.
However, to one of ordinary skill in the art before the effective filing date of the claimed invention, it would have been obvious design choice to incorporate a 2-D or a 1-D color palette. While the combined Chang and Giraud references does not explicitly specify sorting and displaying a 1-D color palette, and generating a two-dimensional (2D) from a one-dimensional (1D), the selection of a 1D vs. 2D layout constitutes as a matter of design choice that does not produce a new or unexpected result (see MPEP § 2144.04 Rearrangement of Parts). A person of ordinary skill in the art would recognize that organizing a palette in 1D or 2D both serve the same purpose to visually present an organized set of colors for selection or analysis. The selection between these formats would have depended on user interface considerations such as available display space, aesthetic preference, or the number of colors to be displayed, falling within the routine skill of the designer. Therefore, it is considered obvious in further view of Chang in view of Giraud, with the limitation being an obvious design choice.
Regarding claim 12, Chang in view of Giraud discloses the computer implemented method of claim 1, and further discloses sorting the colors to output the color palette (Chang Fig. 7-8 and Pg. 243, Section IV, Right Column; Neighboring neurons linked by solid lines represent adjacent entries in the color palette. The absence of long links indicates a good color-ordering palette),
and causing display of the color palette as the 2-D color palette (Chang Fig. 13-14).
Chang in view of Giraud does not appear to explicitly disclose sorting the colors to output the color palette as a 1- D color palette; and generating a 2-D color palette from the 1-D color palette;
Chang and Giraud, along with the motivation to combine the obvious design choice would’ve been the same as that set forth above with respect to claim 10.
Regarding claim 13, Chang discloses one or more non-transitory computer-readable media having a plurality of executable instructions embodied thereon, which, when executed by one or more processors, cause the one or more processors to perform a method (Section IV, Pg. 245, Right Column, Par. 2; All the simulations were run on a Personal Computer equipped with a Pentium Pro. 2.1 GHz Processor and 512 Mbytes of system memory) comprising:
receiving a
(Pg. 239; new image…original color image);
mapping colors to a three-dimensional (3-D) color space by converting the colors from the first format to a format of the 3-D color space and outputting the colors at locations on the 3-D color space (Pg. 239; new image…original color image. Fig. 7 and Pg. 239, Left Column, Step 2); A new image pixel value x(t) = [r(t), g(t), b(t)] ^T);
generating a distribution of points on the 3-D color space by outputting the distribution of points at corresponding locations on the 3-D color space (Embodiment 1 (SOM): Fig. 2 and Pg. 239, Left Column, Step 3); The input x(t) is compared to the weight wi(t) of each neuron i simultaneously, by the square of Euclidean distance in red, green, and blue (RGB) space defined as follows: (1). Embodiment 2 (FS-SOM): Pg. 240, Left Column, 1); initialize the neurons of FS-SOM with uniformly distributed weight vectors on the major diagonal of the input space, i.e. (5) where N is the number of neurons);
organizing via a self-organizing map algorithm, the distribution of points (Embodiment 1 (SOM): Pg. 239, Left Column, Par. 1; Let N be the total number of neurons and wi(t) ∈ R^3 be the weight vector of the ith neuron at epoch time t. Embodiment 2 (FS-SOM): Pg. 240, Left Column, 1); initialize the neurons of FS-SOM with uniformly distributed weight vectors on the major diagonal of the input space, i.e. (5) where N is the number of neurons) to discrete locations on the 3-D color space to output an organized distribution of points on the 3-D color space (Embodiment 1 (SOM): Pg. 239, Left Column, Step 3); the weight wi(t) of each neuron i…in red, green, and blue (RGB) space…Embodiment 2 (FS-SOM): Pg. 240, Left Column, 2); a global butterfly permutation sequence is proposed to present the spatially correlated input data from a multidimensional coordinate system), wherein organizing the distribution of points is based on a relationship between the corresponding locations of the organized distribution of points on the 3-D color space and the locations of the colors on the 3-D color space (Embodiment 1 (SOM): Pg. 239, Left-Right Column; the neurons training to obtain the color palette (Steps 2-5)…Step 3) …The input x(t) is compared to the weight wi(t) of each neuron I simultaneously, by the square of Euclidean distance in red, green, and blue space defined as follows: (1)-(4). Embodiment 2 (FS-SOM): Fig. 7-8 and Pg. 243, Section IV, Right Column; Neighboring neurons linked by solid lines represent adjacent entries in the color palette. The absence of long links indicates a good color-ordering palette…By superimpose the topological maps of Fig. 8 on the color distribution of the original images in Fig. 7);
computing a first set of Euclidian distances between a first location of the locations corresponding to a first color of the colors to the discrete locations of the organized distribution of points (Embodiment 1 (SOM): Pg. 239, Left Column, Step 3); The input tristimulus x(t) is compared to the weight wi(t) of each neuron I simultaneously, by the square of Euclidian distance in red, green, and blue (RGB) space…||x(t) - wi(t)||^2…Embodiment 2 (FS-SOM): Pg. 240, Right Column, 3)-4); adopting the standard SOM best marching unit search with F(ui) = 1…H(i, c, m) is a tapering neighborhood function which is a hallmark of the SOM algorithms. This neighborhood taper alters the update of each neuron, I within the neighborhood of winning neuron, c by a different amount at discrete time interval, m depending on the distance between weight vectors wi and wc…In (7) and (8) the use of a frequency sensitive learning rate function implies that the training time for each individual neuron, rather than the training time of the network, is used to modulate the distance metric for updating);
associating the first color with a first point of the organized distribution points based on a shortest Euclidian distance of the first set of Euclidian distances (Embodiment 1 (SOM): Pg. 239, Left Column, Step 3); The winning neuron c for this input is the one with the smallest distance, i.e., (2)…Right Column, Step 6); For every pixel value in the original color image, the best representative color is ground from the color map based on the same Euclidean distance metric of (1). Embodiment 2 (FS-SOM): Pg. 240, Right Column, 3)-4); adopting the standard SOM best marching unit search with F(ui) = 1…);
computing a second set of Euclidian distances between a second location of the locations corresponding to a second color of the colors to the discrete locations of the organized distribution of points (Embodiment 1 (SOM): Pg. 239, Left Column, Step 3); The input tristimulus x(t) is compared to the weight wi(t) of each neuron I simultaneously, by the square of Euclidian distance in red, green, and blue (RGB) space…||x(t) - wi(t)||^2…Pg. 239, Right Column, Step 4); The epoch time t is increased by one…The process is repeated from Step 2… Embodiment 2 (FS-SOM): Pg. 240, Right Column, 3)-4); adopting the standard SOM best marching unit search with F(ui) = 1…); and
associating the second color with a second point of the organized distribution points (Embodiment 1 (SOM): Pg. 239, Left Column, Step 3); The winning neuron c for this input is the one with the smallest distance, i.e., (2)…Right Column, Step 6); For every pixel value in the original color image, the best representative color is ground from the color map based on the same Euclidean distance metric of (1)... Pg. 239, Right Column, Step 4); The epoch time t is increased by one…The process is repeated from Step 2… Embodiment 2 (FS-SOM): Pg. 240, Right Column, 3)-4); adopting the standard SOM best marching unit search with F(ui) = 1…),
sorting the colors based on the discrete locations, on the 3-D color space, of the organized distribution of points associated with the colors to output the (Fig. 7-8 and Pg. 243, Section IV, Right Column; Neighboring neurons linked by solid lines represent adjacent entries in the color palette. The absence of long links indicates a good color-ordering palette
and causing display of the (Fig. 13-14).
Chang does not disclose receiving a one-dimensional (1-D) color palette comprising a set of colors in a first format, associating the colors with different points of the organized distribution of points, such that no two colors of the colors are associated with a same point of the organized distribution of points, and associating the second color with a second point of the organized distribution points that is not already associated with the first color, output the 1-D color palette comprising the colors as sorted, and causing display of the 1-D color palette comprising the colors as sorted.
In the same art of associating/matching color-representative elements using a nearest-neighbor distance criterion, Giraud discloses associating the colors with different points of the organized distribution of points, such that no two colors of the colors are associated with a same point of the organized distribution of points (Pg. 701, Right Column, Section 2.2, Par. 2; propose to constrain the ANN search and to restrict the number of associations to the same element. To do so, we set a parameter E that defines the maximum number of selection of the same superpixel…E = 1…during the iterative process, a superpixel Ai can be assigned to a superpixel Bk, only if less than E superpixels in A are already assigned to Bk. If Bk is already matched by E elements in A, one superpixel Aj assigned to Bk must be sent to another superpixel in B to allow Ai to match Bk…), and associating the second color with a second point of the organized distribution points that is not already associated with the first color (Pg. 701, Right Column, Section 2.2, Par. 2; propose to constrain the ANN search and to restrict the number of associations to the same element. To do so, we set a parameter E that defines the maximum number of selection of the same superpixel…E = 1…during the iterative process, a superpixel Ai can be assigned to a superpixel Bk, only if less than E superpixels in A are already assigned to Bk. If Bk is already matched by E elements in A, one superpixel Aj assigned to Bk must be sent to another superpixel in B to allow Ai to match Bk…).
It would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to incorporate Giraud’s constraint on match diversity into Chang’s nearest-distance (best-matching) color to node association when forming a palette. Chang’s baseline nearest-neighbor assignment can predictably produce multiple colors mapping to the same representative point that could reduce palette diversity and stable ordering, therefore, applying Giraud’s one-to-one assignment constraint would be a routine optimization in the art of color-space mapping to ensure unique nearest-neighbor assignment, yielding predictable results of better distributed set of representative points for subsequent palette sorting and display.
Chang in view of Giraud does not appear to explicitly disclose receiving a one-dimensional (1-D) color palette comprising a set of colors in a first format, output the 1-D color palette comprising the colors as sorted, and causing display of the 1-D color palette.
However, to one of ordinary skill in the art before the effective filing date of the claimed invention, it would have been obvious design choice to incorporate a 2-D or a 1-D color palette. While the combined Chang and Giraud references does not explicitly specify sorting and displaying a 1-D color palette, and generating a two-dimensional (2D) from a one-dimensional (1D), the selection of a 1D vs. 2D layout constitutes as a matter of design choice that does not produce a new or unexpected result (see MPEP § 2144.04 Rearrangement of Parts). A person of ordinary skill in the art would recognize that organizing a palette in 1D or 2D both serve the same purpose to visually present an organized set of colors for selection or analysis. The selection between these formats would have depended on user interface considerations such as available display space, aesthetic preference, or the number of colors to be displayed, falling within the routine skill of the designer. Therefore, it is considered obvious in further view of Chang in view of Giraud, with the limitation being an obvious design choice.
Regarding claim 14, Chang in view of Giraud and in further view of obvious design choice discloses the one or more non-transitory computer-readable media of claim 13, and further discloses wherein the distribution of points comprises an equal number of points to the colors (Pg. 242, Section III.B, Par. 1; When SOM is used for CQ, the color palette is formed from the final weights wi of the neurons. Hence, the maximum number of colors that can appear in the reconstructed image is equal to the number of neurons N), and wherein associating the colors with different points of the organized distribution of points occurs after organizing the distribution of points to output the organized distribution of points (Pg. 239, Section II; three main phase…the initial network setup (Step 1_, the neurons training to obtain the color palette (Steps 2-5) and the quantizer mapping to generate the pixel indices (Step 6)…Step 5) colormap design has been completed…Step 6) the best representative color is found from the color map based on the same Euclidian distance metric…)
Chang does not disclose such that all points of the organized distribution of points are associated with a different color of the colors after organization via the self-organizing map algorithm.
In the same art of associating/matching color-representative elements using a nearest-neighbor distance criterion, Giraud discloses such that all points of the organized distribution of points are associated with a different color of the colors after organization via the self-organizing map algorithm (Pg. 701, Right Column, Section 2.2, Par. 2; propose to constrain the ANN search and to restrict the number of associations to the same element. To do so, we set a parameter E that defines the maximum number of selection of the same superpixel…E = 1…during the iterative process, a superpixel Ai can be assigned to a superpixel Bk, only if less than E superpixels in A are already assigned to Bk. If Bk is already matched by E elements in A, one superpixel Aj assigned to Bk must be sent to another superpixel in B to allow Ai to match Bk…).
Chang and Giraud, along with the motivation to combine the obvious design choice would’ve been the same as that set forth above with respect to claim 13.
Claim(s) 15 is/are rejected under 35 U.S.C. 103 as being unpatentable over Chang et al. "New adaptive color quantization method based on self-organizing maps." IEEE transactions on neural networks 16, no. 1 (2005): 237-249, hereinafter referred to as “Chang”, in view of Giraud et al. "Superpixel-based color transfer." In 2017 IEEE International Conference on Image Processing (ICIP), pp. 700-704. IEEE, 2017, hereinafter referred to as “Giraud”, and in further view of obvious design choice (MPEP § 2144.04), and in further view of Zhang et al. "Color clustering using self-organizing maps." In 2007 International Conference on Wavelet Analysis and Pattern Recognition, vol. 3, pp. 986-989. IEEE, 2007, hereinafter referred to as “Zhang”.
Regarding claim 15, Chang in view of Giraud and in further view of obvious design choice discloses the one or more non-transitory computer-readable media of claim 13, and further discloses wherein associating the colors with the different points of the organized distribution of points occurs after organizing the distribution of points to output the organized distribution of points (Chang Pg. 239, Section II; three main phase…the initial network setup (Step 1_, the neurons training to obtain the color palette (Steps 2-5) and the quantizer mapping to generate the pixel indices (Step 6)…Step 5) colormap design has been completed…Step 6)the best representative color is found from the color map based on the same Euclidian distance metric…).
Chang does not disclose wherein the distribution of points comprises a number of points that is greater than the colors, and each point in the organized distribution of points that is not associated with any color of the colors is not used for displaying the colors as sorted, and such that all of the colors are associated with a different point of a subset of the organized distribution of points after organization via the self-organizing map algorithm.
In the same art of associating/matching color-representative elements using a nearest-neighbor distance criterion, Giraud discloses such that all points of the organized distribution of points are associated with a different color of the colors after organization via the self-organizing map algorithm (Pg. 701, Right Column, Section 2.2, Par. 2; propose to constrain the ANN search and to restrict the number of associations to the same element. To do so, we set a parameter E that defines the maximum number of selection of the same superpixel…E = 1…during the iterative process, a superpixel Ai can be assigned to a superpixel Bk, only if less than E superpixels in A are already assigned to Bk. If Bk is already matched by E elements in A, one superpixel Aj assigned to Bk must be sent to another superpixel in B to allow Ai to match Bk…).
Chang and Giraud, along with the motivation to combine the obvious design choice would’ve been the same as that set forth above with respect to claim 13.
Chang in view of Giraud does not disclose wherein the distribution of points comprises a number of points that is greater than the colors, and each point in the organized distribution of points that is not associated with any color of the colors is not used for displaying the colors as sorted.
In the same art of color-palette generation/ordering via color space clustering and nearest neighbor assignment, Zhang discloses wherein the distribution of points comprises a number of points that is greater than the colors (Pg. 988, Section 3, Right column; set the size of SOM to 16*16…The image colors are reiteratively used to train the network…each color point is cyclically chosen from the data set, and presented to all neurons on the map simultaneously. Examiner's note: Because each color point is presented to all neurons simultaneously, for any given color point, there exists a plurality of neurons available to be matched, therefore there are more neurons (points) than colors (input color point)), and each point in the organized distribution of points that is not associated with any color of the colors is not used for displaying of the colors as sorted (Pg. 988, Section 3, Right column and Fig. 3; The next color point is presented to the network at time t+1. The new winning neuron is chosen by repeating the procedure from step 2 until all iterations have been made. Examiner's note: only the winning neurons are associated with a color in the set of colors, and used for displaying in Fig. 3).
It would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to incorporate Zhang’s teaching of configuring the SOM with more neurons (points) than input colors and only displaying selected entries into Chang and Giraud’s SOM-based palette construction system. A larger codebook is a routine way to increase representative resolution and reduce quantization error. When Chang and Giraud’s combined system has match-diversity constraint that prevents reuse of the same representative, a larger codebook of candidate neurons is a routine way to increase representative resolution, allowing the system to select a near-optimal alternative rather than forcing multiple inputs into the same representative or accepting a poor match, providing predictable results in improved reconstruction/palette fidelity and output consistency across varying palettes/distributions. Further, displaying unassigned points provides no color information and would unnecessarily clutter the palette output/display, therefore the combination is additionally a predictable UI optimization to omit any points to reduce visual clutter and improve user experience.
Claim(s) 16 is/are rejected under 35 U.S.C. 103 as being unpatentable over Chang et al. "New adaptive color quantization method based on self-organizing maps." IEEE transactions on neural networks 16, no. 1 (2005): 237-249, hereinafter referred to as “Chang”, in view of Giraud et al. "Superpixel-based color transfer." In 2017 IEEE International Conference on Image Processing (ICIP), pp. 700-704. IEEE, 2017, hereinafter referred to as “Giraud”, in further view of obvious design choice (MPEP § 2144.04), and in further view of Xiao et al. “Self-organizing map-based color palette for high-dynamic range texture compression”. Neural Comput & Applic 21, 639-647 (2012), pages 640-647, hereinafter referred to as “Xiao”
Regarding claim 16, Chang in view of Giraud, and in further view of obvious design choice discloses the one or more non-transitory computer-readable media of claim 13, but does not disclose wherein causing display of the 1-D color palette further comprises:
causing display of an adjustable parameter, wherein the adjustable parameter generates linear-interpolated colors by linear-interpolating between the colors as sorted and corresponding colors represented by the discrete locations of the organized distribution of points associated with the colors; receiving a change to the adjustable parameter; and causing display of the linear-interpolated colors based on the change to the adjustable parameter.
In the same art of SOM-based training for color organization, Xiao discloses wherein causing display of the color palette further comprises:
causing display of an adjustable parameter (Pg. 642, Section 3, Formulas (6) and (7) definition of p and q as normalized coordinates. Examiner's note: p and q are adjustable parameters because they determine interpolation weights, adjusting p and q changes the displayed colors as seen in Figure 4), wherein the adjustable parameter generates linear-interpolated colors by linear-interpolating between the colors as sorted and corresponding colors represented by the discrete locations of the organized distribution of points associated with the colors (Pg. 642, Section 3; perform linear interpolation along the horizontal direction to get two intermediate color values, given by Formula (8) and (9)…linear interpolation along the vertical direction, given by Formula (10));
receiving a change to the adjustable parameter (Formulas (6) and (7) p and q. Examiner's note: changing p and q changes which interpolated value is computed, Fig. 3b and 4 showing how changing coordinates values (and thus p and q) leads to new interpolated palette colors);
and causing display of the linear-interpolated colors based on the change to the adjustable parameter (Fig. 4 shows a 16x16 palette expanded to a 256x256 palette by interpolation)
It would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to combine Xiao’s linear interpolation method for virtual palettes with Chang and Giraud’s SOM-based color sorting technique. Doing so provides the predictable benefit of smoother transitions and high-resolution palettes, with interpolation being a well-known and common technique in image processing.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JENNY NGAN TRAN whose telephone number is (571)272-6888. The examiner can normally be reached Mon-Thurs 8am-5pm.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Alicia Harrington can be reached at (571) 272-2330. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/JENNY N TRAN/Examiner, Art Unit 2615
/ALICIA M HARRINGTON/Supervisory Patent Examiner, Art Unit 2615